What New Zealand looked like during the last ice age

Stand on Wellington’s south coast today and the sea starts at your feet. Stand on the same spot 21,000 years ago and you would have looked out over dry land running away to the south for tens of kilometres, with the coast a long walk away and no strait at all between you and the South Island.

That is the world of the Last Glacial Maximum, the coldest, driest, iciest moment of the last ice age. So much of the planet’s water was stacked up on land as ice that the sea fell more than 100 metres, and coastlines everywhere retreated to the edge of the continental shelf. New Zealand was not three main islands. It was one.

We have built an interactive globe that shows it. The Ice Age Globe lets you spin the Earth and switch between seven moments in time, from the peak of the ice 21,000 years ago through to today, watching the great ice sheets grow and collapse while the coastlines move. It is built from the same reconstruction that geophysicists use to study how the Earth’s crust is still rebounding from the weight of that vanished ice.

The short answer

At the peak of the last ice age, around 21,000 years ago, New Zealand’s North and South Islands were joined into a single landmass by a plain west of today’s Cook Strait. The world was about 6 degrees colder, sea level was more than 100 metres lower, and roughly two and a half times as much ice sat on land as sits there now. Glaciers filled the Southern Alps, but most of the country was ice free grassland and scrub. No people lived here, and none would for another 18,000 years.

Why ice ages happen

Before the map, the mechanism. Ice ages are not random cold spells. They run on a clock, and the clock is orbital.

The Earth’s path around the Sun is not perfectly fixed. Its orbit stretches from more circular to more elliptical and back over roughly 100,000 years. The tilt of its axis, which is what gives us seasons, wobbles between about 22.1 and 24.5 degrees over 41,000 years. And the axis itself precesses, swivelling like a slowing top over about 23,000 years, changing which season falls at the closest approach to the Sun. These three cycles were worked out in detail by the Serbian mathematician Milutin Milanković in the 1920s and now carry his name.

None of them changes the total amount of sunlight the Earth receives by very much. What they change is how it is distributed, particularly how much falls on the northern high latitudes in summer. That turns out to be the switch. Snow that survives one summer is snow that can accumulate the next, and once an ice sheet begins to grow it helps itself along, because bright ice reflects sunlight that dark ground would have absorbed.

Feedbacks do the rest. Colder oceans absorb more carbon dioxide, which cools the planet further. Drier air carries less water vapour, another greenhouse gas. The result is that a modest nudge in the distribution of sunlight can swing the whole system between a glacial world and an interglacial one, and the swings have been running on that rhythm for the last couple of million years. The Last Glacial Maximum is simply the most recent trough, and the one we can see most clearly.

The world at the Last Glacial Maximum

The Last Glacial Maximum is the name geologists give to the period roughly 26,500 to 19,000 years ago, when the ice sheets of the last glacial period reached their greatest size. The single moment usually used as the peak is about 21,000 years ago, and that is where our globe opens.

It was a different planet, though not an unrecognisable one. One widely cited 2020 reconstruction, led by Jessica Tierney and published in Nature, put global average temperature about 6 degrees Celsius below the modern average, with other estimates spanning roughly 4 to 7 degrees. That sounds modest until you remember it is a global average, smeared across oceans that change slowly. Over the northern continents the difference was far larger, and it was enough to bury Canada and Scandinavia under kilometres of ice.

Carbon dioxide in the atmosphere sat around 190 parts per million, measured directly from air bubbles trapped in Antarctic ice cores. Today it is above 420. The air was also drier and much dustier, because cold air holds less moisture and because so much bare ground lay exposed at the edges of the ice.

The ice itself is the headline. Processing the ICE-6G_C reconstruction ourselves gives a grounded ice volume at 21,000 years ago of about 75 million cubic kilometres, two and a half times the roughly 29 million that exists on Earth today. All of that water came out of the ocean, and the ocean noticed.

Where the sea went

An ice sheet is, in the plainest terms, ocean water that fell as snow and never went home. Every cubic kilometre locked into the Laurentide or Fennoscandian ice sheets was a cubic kilometre missing from the sea.

Published estimates put global sea level at the Last Glacial Maximum roughly 120 to 130 metres below today’s. Our own calculation from the model’s ice volumes lands slightly under that, at about 117 metres, because the method we use is a simple one that spreads the meltwater over an ocean whose exact area and depth changed as the coastlines moved. Either way the effect on the map is dramatic, because most continents are fringed by a shallow shelf that a 100 metre drop turns into dry land.

An important subtlety, and one the globe shows well, is that the sea did not fall by the same amount everywhere. The weight of an ice sheet pushes the crust down beneath it and makes it bulge up around the edges, and the gravitational pull of that much ice actually heaps seawater towards it. Add the slow flow of rock in the mantle and the result is that local sea level change can differ from the global average by tens of metres. Near New Zealand the model puts the fall at a bit over 100 metres rather than the global 120.

New Zealand as one island

This is the part that surprises most New Zealanders. Drop the sea by 100 metres and the country stops being an archipelago.

The seafloor between the North and South Islands is not uniformly deep. Cook Strait itself contains a genuinely deep trench, a submarine canyon system that drops away hundreds of metres, but to the west the South Taranaki Bight is shallow, a broad shelf between Taranaki and the top of the South Island. At ice age sea levels that shelf was a plain, and it joined the two main islands into a single landmass. In our processing of the model, roughly 70 percent of the strait region stands above water at the peak, falling to 40 percent by 14,500 years ago and 18 percent by 13,000 years ago as the sea came back through. Stewart Island joined the South Island too.

What we now call Cook Strait was not a strait at all. With the western plain dry, the deep water east of it became a long inlet of the Pacific, reaching in from the east and stopping. You could have walked from Wellington to Nelson, though you would have gone the long way round, well to the west of the ferry route.

Around the rest of the coast the same shelf logic applied. The Hauraki Gulf was a coastal plain with rivers running across it. Both of Auckland’s harbours were dry valleys. Ninety Mile Beach sat well inland of the actual shore. Off the east coast of the South Island the Canterbury Bight extended much further out, and the Chatham Rise, which today lies a few hundred metres below the surface, remained submerged but shallower.

Nobody saw any of it. This matters, and it is worth being blunt about, because ice age land bridges elsewhere in the world are bound up with the story of human migration. New Zealand’s are not. The first people arrived here around 1280 AD, according to the radiocarbon work led by Janet Wilmshurst and published in PNAS in 2008, which dated both Pacific rat bones and distinctive rat-gnawed seeds. That is roughly 18,000 years after the ice age peak. The plain that joined the North and South Islands had been under water for more than ten thousand years before the first waka arrived. Lower seas connected New Zealand’s islands to each other, never to anywhere else, and the nearest continental shelf still ended in thousands of kilometres of open ocean.

Ice in the Southern Alps

New Zealand had real glaciers, and they did real work. Ice filled the valleys of the Southern Alps and pushed out onto the plains on both sides. In the south it reached the lowlands, gouging out the long finger lakes of Otago and Southland. Fiordland’s fiords are glacial valleys, cut by ice and then flooded by the returning sea.

The timing here is genuinely interesting, because New Zealand did not simply follow the northern hemisphere. Work led by Henrik Rother and James Shulmeister, published in PNAS in 2014 and titled “The early rise and late demise of New Zealand’s last glacial maximum”, dated moraines in the Rangitata Valley and found the local glacial maximum happened shortly before 28,000 years ago, ahead of the global peak, followed by a series of successively smaller readvances between about 26,000 and 19,000 years ago. The retreat afterwards was gradual rather than a collapse. The biggest advance did not line up with the coldest temperatures, which tells you that precipitation and wind patterns, not temperature alone, drive mountain glaciers.

One honest caveat about our globe. The ICE-6G_C model works on a grid of one degree, roughly 110 kilometres across. That is fine for ice sheets the size of continents. It is far too coarse to draw an ice field that fits inside a mountain range, so New Zealand looks almost bare on the globe even at the peak of the ice. The glaciers were there. The model simply cannot resolve them, and no amount of zooming will make them appear.

The New Zealand climate of the ice age

Cold is the obvious difference, and estimates for New Zealand at the glacial peak generally sit somewhere around 4 to 6 degrees below present. But temperature is the least interesting part of the change.

It was windier. The belt of westerly winds that dominates New Zealand’s weather shifted and strengthened, and with a bigger, drier landmass and less forest to slow it down, the wind had a freer run. It was dustier for the same reasons, and layers of wind blown silt from this period, called loess, blanket parts of Canterbury, Otago, Wairarapa and Manawatū to depths of several metres. If you have seen a pale, fine grained bank in a road cutting in the South Island high country, you have probably looked at ice age dust.

It was also, in places, drier, and the vegetation reflected it. Instead of the near continuous forest that greeted the first arrivals, much of the country carried grassland, herbfield and low shrubland, with forest holding on in warmer, wetter refuges in the north and in sheltered valleys. The treeline was hundreds of metres lower than today. Frost reached much further north.

The animals were the ones New Zealand is famous for, and they had the place entirely to themselves. Nine species of moa, the giant flightless birds, ranged across the open country, hunted from above by Haast’s eagle, the largest eagle known to have existed. There were no land mammals apart from bats. That ecosystem survived the glacial peak, the warming, the sea level rise and the forest’s return without much trouble. What it did not survive was human arrival 18,000 years later, which is a different story with a much shorter timescale.

The eruption that dates the ice age

New Zealand contributed something else to ice age science, and it came out of the ground near Taupō.

The Ōruanui eruption was the largest volcanic event anywhere on Earth in the last 70,000 years. It emptied a magma chamber under what is now Lake Taupō, collapsed the caldera that the lake sits in, and threw ash across the whole country and far out into the Pacific. Radiocarbon work led by Marcus Vandergoes dates it to about 25,580 years before present, give or take a couple of centuries, which places it just before the peak of the last ice age rather than in the middle of it.

That precise date turned the ash layer into a scientific gift. A volcanic ash layer is laid down everywhere at effectively the same instant, so wherever you find it you have a fixed point in time. The Kawakawa-Ōruanui tephra shows up in lake beds and swamps across New Zealand, in seafloor sediment cores, and, remarkably, in Antarctic ice. A 2017 study in Scientific Reports identified volcanic glass from the eruption in an Antarctic ice core, giving researchers a shared time marker to line up New Zealand’s climate records with Antarctica’s during the Last Glacial Maximum.

The northern ice sheets

Spin the globe north and the scale of the thing becomes obvious. Two enormous ice sheets did most of the work of lowering the world’s oceans.

The Laurentide ice sheet covered almost all of Canada and pushed well down into the United States, its southern edge running roughly through the sites of New York, Chicago and Seattle. In our processing of the model it exceeds 3,700 metres thick over parts of the Canadian Shield. To its west the smaller Cordilleran ice sheet filled British Columbia and the mountains behind it, and where the two met along the eastern front of the Rockies the combined ice was thicker still.

Across the Atlantic, the Fennoscandian ice sheet buried Scandinavia, Finland, the Baltic states and a good part of northern Russia, reaching a peak thickness in the model of about 2,700 metres over the Gulf of Bothnia. Its southern margin ran through what is now northern Germany and Poland. A separate British-Irish ice sheet covered Ireland, Scotland, Wales and northern England, and at times joined the Fennoscandian sheet across the northern North Sea.

In the southern hemisphere the Antarctic ice sheet was larger than today, grounded out across the continental shelf, and a Patagonian ice sheet ran down the spine of the southern Andes. But the southern hemisphere is mostly ocean, and ocean does not grow ice sheets. The great asymmetry of the last ice age is that the north did nearly all of the freezing while the whole world paid for it in sea level.

Ice thickness over seven cities

Numbers like “three kilometres of ice” are hard to feel. So we sampled the model at seven cities and drew the result against their skylines, which is what the city buttons on the globe do. The figures below are model values for the grid cell containing each city, so they are good estimates rather than survey measurements, and they change as you move through the eras.

Wellington and Auckland. No ice at all, at any point in the record. The New Zealand story is entirely about the sea. Auckland at the peak sat about 64 kilometres from open ocean in our processing of the model, against 9 kilometres today, with the Hauraki Gulf a plain and both harbours dry valleys. Wellington’s figure is smaller, around 13 kilometres against 3 today, because the deep trench east of the city stayed flooded even at the lowest sea levels. That is the point about Cook Strait in miniature. The deep part remained water. The shallow plain to the west did not. This is why the New Zealand entries on the globe show a coastline distance instead of an ice wall.

Chicago. Under about 515 metres of ice at the peak, sitting near the southern margin of the Laurentide sheet. There was no Lake Michigan. The Great Lakes are a product of this ice, though not quite in the way people often say. The basins follow river valleys and soft rock that were already there. The ice deepened and scoured them, then filled them with meltwater as it pulled back.

New York. Close to the edge, and the most uncertain figure of the seven. The model puts about 780 metres of ice over the city’s grid cell at the peak, but Manhattan sat near the southern limit of the Laurentide sheet, and Long Island is built from the debris the glacier dumped at its front, the Ronkonkoma moraine marking the furthest advance and the Harbor Hill moraine a later pause. An ice margin is a line rather than an area, and a model on a 110 kilometre grid cannot tell you which side of that line a given street was on. Treat New York as a margin city where the exact number matters less than the fact that the ice ended right about there.

Calgary. Buried under about 1,790 metres of ice at the peak, in the zone where the Laurentide and Cordilleran sheets met. As they separated during the melt, an ice free corridor opened along this line, sometime around 15,000 to 14,000 years ago. It used to be the standard explanation for how the first people reached the Americas. It is now contested, with a coastal route along the Pacific margin looking more likely for the earliest arrivals, and the corridor treated as one possible route among several.

Stockholm. About 1,630 metres of ice at the peak in our sampling, with the thickest part of the Fennoscandian dome, close to 2.7 kilometres, sitting further north over the Gulf of Bothnia. It is also the only one of the seven cities still under ice late in the record, with more than 800 metres above it as recently as 13,000 years ago. There was no Baltic Sea, only ice. Stockholm’s most striking ice age legacy is not visible at all, it is measurable. The crust is still rising as it rebounds from a load that vanished 10,000 years ago, at a few millimetres a year around Stockholm and faster further north.

London. Ice free, and that is the point of including it. The British-Irish ice sheet stopped a few hundred kilometres to the north. London sat on treeless periglacial steppe, frozen hard for much of the year, grazed by mammoth, reindeer and horse. Britain was not an island. The Thames was a tributary of the Rhine, and the two rivers ran together across the dry floor of the North Sea.

The land that the low sea uncovered

Switch the globe to 21,000 years ago and look at the tan areas. That is land that exists on no modern map, and there was a lot of it. By our count from the model, about 26 million square kilometres of what is now seabed stood above water at the peak, an area larger than South America. By 8,000 years ago it is down to about 1.4 million, and today it is gone.

Beringia is the famous one, a broad plain joining Siberia to Alaska where the Bering Strait is now. It was not a narrow bridge but a subcontinent, up to 1,000 kilometres from north to south, cold and dry and largely unglaciated. Mammoth, bison and horse lived there, and people crossed it.

Sundaland joined Sumatra, Java, Borneo and the Malay Peninsula into one continuous extension of Asia. To its south east, Sahul joined Australia, New Guinea and Tasmania into a single continent. The two were never connected, and the deep water between them had to be crossed by boat, which is why the settlement of Australia at least 50,000 years ago is such a significant moment in the history of seafaring.

Doggerland is the North Sea floor, a landscape of rivers, marsh and woodland between Britain and Denmark, well populated by Mesolithic hunters whose tools and bones still come up in fishing nets. It is the best studied drowned country in the world, and the most poignant, because we can date its disappearance fairly closely.

Smaller versions of the same thing appeared everywhere. The Persian Gulf, which is shallow, was a river valley rather than a sea. Sri Lanka was joined to India, Japan very nearly joined to the Asian mainland, and the Falkland Islands sat on a much broader Patagonian shelf. Tasmania was part of Australia. The Mediterranean, being deep, changed comparatively little, which is why its coastline on the globe looks reassuringly familiar while the North Sea and South East Asia look like different planets.

All of this drowned land is the reason coastal archaeology is so difficult. For most of human history, people lived on coasts, and for most of the period archaeologists care about, those coasts are now under water. The record of how people lived at the ice age shoreline is largely offshore, buried in sediment, and only occasionally dragged up by a trawler.

The animals of the mammoth steppe

The tan land on the globe was not empty. The vast unglaciated belt running from western Europe across Siberia and over Beringia into Alaska carried an ecosystem with no modern equivalent, usually called the mammoth steppe. It was cold, dry, treeless and, crucially, productive, a grassland that supported grazing animals in enormous numbers.

Woolly mammoth is the emblem, but the steppe carried woolly rhinoceros, steppe bison, wild horse, reindeer, musk ox, saiga antelope and cave lion. It stretched across a third of the northern hemisphere’s land, and it worked because the cold, dry air kept the ground clear of the forest and bog that dominate those latitudes today.

Almost all of it is gone, and it went as the ice age ended rather than at its peak. The warming brought moisture, the moisture brought forest and peat, the grassland fragmented, and the large grazers went with it. Human hunting was part of the story, and researchers still argue about how large a part, but the loss of the habitat itself is the change the globe illustrates. Watch Beringia disappear beneath the sea between the 14,500 and 11,500 year eras and you are watching an entire biome’s core territory drown.

Nothing remotely like this existed in New Zealand, which had no large mammals to lose and whose big animals were birds. It is a useful contrast to keep in mind while spinning the globe. The northern hemisphere’s ice age story is about ice and megafauna. The southern hemisphere’s, and certainly New Zealand’s, is mostly about water.

How the ice melted

The melt was not a steady thaw. It came in surges and stalls, and the eras on the globe are chosen to catch them. These are the figures our processing of the model produces for each one.

Era	Sea level vs today	Ice vs today	Extra dry land
21,000 years ago	117 m lower	2.6 times	25.8 m km²
14,500 years ago	92 m lower	2.2 times	21.4 m km²
13,000 years ago	58 m lower	1.8 times	15.0 m km²
11,500 years ago	42 m lower	1.6 times	11.7 m km²
8,000 years ago	6 m lower	1.1 times	1.4 m km²
6,000 years ago	about today’s	1.0 times	0.4 m km²
Today	baseline	baseline	none

Two things stand out in that table. The first is how front loaded the melt was, with more than half the sea level rise done by 13,000 years ago. The second is the collapse of the tan land. Between 11,500 and 8,000 years ago the world loses more than 10 million square kilometres of coastal plain, which is the period when Doggerland, Beringia and Sundaland all go under.

14,500 years ago, the fastest rise on record

After tens of thousands of years of cold, the northern hemisphere warmed sharply at what is called the Bølling transition, and the ice sheets let go. The result was Meltwater Pulse 1A, the fastest sustained sea level rise in the geological record. Work led by Pierre Deschamps on drowned coral reefs off Tahiti, published in Nature in 2012, dated it to between about 14,650 and 14,310 years ago and put the rise at 14 to 18 metres in under 340 years, a rate exceeding 40 millimetres a year.

For comparison, global sea level is currently rising at a bit over 4 millimetres a year. Meltwater Pulse 1A was roughly ten times that, sustained for three centuries, and it happened when there was an enormous volume of vulnerable ice sitting at low latitudes ready to go. The comparison is worth making carefully in both directions. The ice age world had far more ice to lose than ours does, so a repeat of that exact rate is not on the table. What the episode does establish is that ice sheets are capable of collapsing much faster than a gradual thaw.

13,000 years ago, the cold snap

Then it stopped. Between roughly 12,900 and 11,700 years ago, the northern hemisphere fell back into near glacial cold in a matter of decades, an episode named the Younger Dryas after an Arctic wildflower whose pollen shows up in European sediments of the period. Glaciers readvanced. The North Atlantic froze further south. The leading explanation is that a vast pulse of fresh meltwater into the North Atlantic disrupted the ocean circulation that carries heat north, though the details are still argued over.

The Younger Dryas is the most important thing on this globe for understanding climate, because it is the clearest evidence in the record that the climate system does not always change gently. It can flip.

11,500 years ago, the Holocene begins

The Younger Dryas ended as abruptly as it began, and the warmth stuck. Geologists mark this as the start of the Holocene, the interglacial we are still in. The Fennoscandian ice sheet retreated off the Baltic, which spent the next few thousand years cycling between a freshwater lake and an arm of the sea as the ice dams and the rising ocean fought over it.

8,000 years ago, the last remnant

By now most of the ice is gone. What survives on the globe is a stubborn remnant of the Laurentide over northern Quebec and Labrador, more than a kilometre thick even at this late stage, the last of an ice sheet that once covered a continent.

Around this time the sea finished Doggerland. About 8,150 years ago a colossal submarine landslide off the Norwegian coast, the Storegga Slide, sent a tsunami across the North Sea. It was long assumed to have drowned Doggerland in a day. Sediment core work published in Antiquity in 2020 revised that. Much of Doggerland was already under water, and the higher ground of the Dogger Bank survived the wave as an archipelago for centuries afterwards, finally drowned not by a catastrophe but by the ordinary, relentless rise of the sea.

6,000 years ago, modern coastlines

Sea level rise slows to a crawl and the world’s coastlines settle into shapes we would recognise. In New Zealand this is roughly when the harbours and estuaries reached their modern form, and when the coastal landscape that Māori would eventually settle finished taking shape, several thousand years before anyone arrived to see it.

Reading the ice age in the landscape today

The best thing about all of this is that you do not need the globe to see it. The ice age is written across New Zealand in features most people drive past without a second look.

The southern lakes. Wakatipu, Wānaka, Hāwea, Te Anau, Manapōuri and Ōhau are all glacial. Their long, narrow, deep shapes are valleys that glaciers gouged out, dammed at the downstream end by the piles of rubble the ice left when it stopped. Wakatipu’s famous dog leg shape follows the path the ice took.

The Canterbury Plains. Not a floodplain in the ordinary sense but a vast apron of gravel, washed out from the mountains by meltwater rivers over successive glaciations and spread across the coast. The braided rivers that cross it are the surviving remnant of a much larger meltwater system.

The fiords. Milford, Doubtful and the rest are glacial valleys cut below the level of the sea at the time, then flooded when the water came back. The characteristic shape, a steep sided U rather than a river’s V, is the signature of ice rather than water.

Every harbour in the country. Waitematā, Manukau, Whangārei, Tauranga, Otago, Wellington and the Marlborough Sounds are drowned river valleys, cut when the sea was low and flooded when it rose. That is why they branch like rivers, because they were rivers.

The loess banks. Those pale, fine, almost powdery cuttings in the Canterbury and Otago hill country are compacted ice age dust, blown off the dry riverbeds and exposed plains when the wind was stronger and the land was bare.

How scientists rebuilt the ice age world

Nobody measured any of this at the time, so where do the numbers come from?

It is a fair question, and the honest answer is that an ice age reconstruction is not a photograph. It is a model, fitted to several independent kinds of evidence that each capture part of the picture, and it is only as good as the agreement between them.

Partly from direct evidence. Moraines mark where a glacier stopped, and cosmogenic dating can tell you when a boulder was first exposed to the sky. Drowned coral reefs record former sea levels. Pollen in lake beds records the vegetation, and therefore the climate. Ice cores give temperature and atmospheric composition directly.

Each of those has a blind spot. Moraines tell you where ice stopped but not how thick it was behind that line. Corals give you sea level at one point on the globe, which is not the global average. Ice cores are superb but come from two ice sheets, so they tell you a great deal about Greenland and Antarctica and comparatively little about Nebraska. Building a global picture means reconciling all of them and accepting that some regions are much better constrained than others.

The clever part is the rebound. The Earth’s mantle behaves over long timescales like an extremely stiff fluid, so an ice sheet pressed the crust down beneath it, and when the ice melted the crust began to rise again. It is still rising. That ongoing motion, called glacial isostatic adjustment, is measurable today with GPS to a fraction of a millimetre a year, and it is effectively a fingerprint of how much ice sat where and when it left.

The reconstruction behind our globe, ICE-6G_C (VM5a), was built by W.R. Peltier, Donald Argus and Rosemarie Drummond at the University of Toronto by tuning an ice history until it simultaneously matched present day GPS uplift measurements, the record of past sea levels around the world, and the dated evidence for where ice margins stood. The result is a set of global grids of ice thickness and topography at 500 year intervals back to 26,000 years ago, and it is freely published.

For the globe we took seven of those time slices, worked out how much the bedrock itself has moved since the ice age at each point, added that to the NOAA ETOPO 2022 bedrock relief model, and drew each era’s coastline at the zero metre line. Basins that end up below sea level but cut off from the ocean are drawn as lakes rather than sea. Then we rendered it onto a sphere. The whole method is deliberately simple and reproducible, and the limitations are the ones already noted, chiefly that a one degree grid cannot see mountain glaciers.

What happened to New Zealand afterwards

The country the ice age left behind was not the one that exists now. As the sea rose it flooded the plain between the islands, opened Cook Strait, drowned the river valleys that became Auckland’s harbours and the Marlborough Sounds, and cut Stewart Island loose.

The vegetation changed just as completely. At the glacial peak much of the country was grassland, shrubland and cold scrub, with forest confined to pockets. As the climate warmed, forest expanded until it covered something like 80 percent of the land area, the beech and podocarp forests that met the first arrivals. That forest is the direct product of the post glacial warming, and it is the one ice age legacy New Zealanders interact with daily.

The other legacy is the shape of the coast itself. Every drowned valley on the New Zealand map, and there are hundreds, was cut by a river when the sea was low and flooded when it rose. Sediment has been filling them in ever since, which is why the upper reaches of harbours like the Waitematā and Tauranga are mudflats rather than deep water, and why estuaries need dredging. That process has been running for around 7,000 years and has not stopped. The tides that shift the sediment around today are working on a coastline the ice age built, and you can watch them run in our New Zealand tide map.

Two other things worth knowing. The ice age is not over. We are in an interglacial, a warm interval within an ice age that began about 2.6 million years ago, and the proof is that Antarctica and Greenland still carry continental ice sheets holding roughly 65 metres of potential sea level rise between them. And the land is still moving. Parts of New Zealand are rising and others sinking, driven mostly by our own tectonics rather than ice rebound, which is one reason local sea level change here differs from the global figure.

What the ice age can and cannot tell us about now

A piece like this invariably raises the modern comparison, so it is worth being careful about what the deglaciation does and does not demonstrate.

What it does not show is a like for like precedent. The ice age world had a colossal amount of ice sitting at temperate latitudes, in places like Chicago and Copenhagen, which was always going to melt once the orbital forcing turned. Nothing comparable exists now. The ice that remains is concentrated in Antarctica and Greenland, in far colder places, and it behaves differently.

What it does show is how the system responds when it is pushed. Three findings from this record are relevant and well established. Ice sheets can retreat far faster than they build, taking tens of thousands of years to grow and a few thousand to disappear. Sea level responds to ice volume without much delay, and Meltwater Pulse 1A shows that response can run at metres per century when a large ice sheet destabilises. And the Younger Dryas shows that the climate can reorganise abruptly, in decades rather than centuries, when ocean circulation is disturbed.

The other thing the record gives you is a sense of proportion about coastlines. New Zealanders tend to treat the shoreline as a permanent feature, the fixed edge of the map. On the timescale of this globe it is nothing of the sort. It has moved tens of kilometres in living geological memory, and every harbour and estuary in the country is a drowned valley, still filling in.

Using the Ice Age Globe

Drag to spin, scroll or pinch to zoom. The buttons along the bottom switch between the seven eras, and the globe crossfades from one to the next so you can watch the ice and the coastline move.

White is grounded ice, shaded by thickness so the domes stand out from the thin margins. Green and brown is land that still exists. Tan is land that only existed because the sea was lower, which is the layer worth watching as you step forward through time. The thin orange line is today’s coastline, drawn on every era so you always have a reference.

Tap a city marker for its ice thickness at that moment, drawn against its skyline at the same scale, or for a coastline comparison where there was no ice. The stats above the globe give the sea level and total ice volume for the era you are looking at.

Sources and data. Ice thickness and palaeotopography from ICE-6G_C (VM5a) by W.R. Peltier, D.F. Argus and R. Drummond, University of Toronto, described in Journal of Geophysical Research (2015) and Geophysical Journal International (2014). Modern relief from NOAA ETOPO 2022. Processing and rendering by Newswire.

Have you found something on the globe that surprised you, or do you know a stretch of coast whose ice age shape you can still read in the landscape? Tell us in the comments below.

This article was written by AI, briefed to report the facts, hopefully without some of the bias people bring to the job 🙂